The Optical Trapezoidal method is a recently proposed method using solely optical remote sensing data with promising potential in hydroclimatic applications; however, its application in evapotranspiration (ET) partitioning is still questionable. This study therefore improved an Optical Trapezoidal-based Evaporative Fraction (OPTREF) model reliant on the physical association between soil surface properties and shortwave infrared transformed reflectance (STR). The proposed model was parameterized using trapezoidal shapes formed by STR and Normalized Difference Vegetation Index (NDVI) retrieved from Sentinel-2 to compute contribution of transpiration to ET (T/ET) in the Central Valley region of California, United States. Results indicated that the OPTREF model efficiently captured regional plant characteristics and seasonal variation of T/ET for three typical land cover types in study region (cropland, grassland, and savanna), particularly T/ET discrepancies rising from alternative cropping systems in cropland. Importantly, predicted T/ET values were in good agreement with previous modelling and empirical reports; for example, T/ET ranges in [0.37-0.75] for cropland, [0.30-0.82] for grassland, and [0.39-0.77] for savanna. In addition, applicability of the OPTREF approach was investigated in tracking agricultural drought using evaporative drought index (EDI), together with analysis of physiological response of T/ET to drought. The OPTREF-based EDI consistently reflected the same drought patterns indicated by records of vegetation health index and drought area percentages data from NOAA. Also, our findings uncovered that under extreme drought, the T/ET ratio is likely to increase and might be approximately equal to 1.0 in areas covered by drought-tolerant plants. Overall, this OPTREF approach offers novelty, simplicity, and versatility to satellite-based remote sensing of ET partitioning and identifying agricultural drought, especially in the context of rapid climate variation.

光学梯形法是最近提出的一种仅使用光学遥感数据的方法，在水文气候应用中具有广阔的潜力；然而，它在蒸发蒸腾 ( ET ) 分区中的应用仍然值得怀疑。因此，本研究改进了基于光学梯形的蒸发分数 (OPTREF) 模型，该模型依赖于土壤表面特性与短波红外转换反射率 ( STR )之间的物理关联。所提出的模型使用由STR形成的梯形形状和从 Sentinel-2 检索的归一化差异植被指数 ( NDVI ) 进行参数化，以计算蒸腾对ET的贡献( T/ET) 在美国加利福尼亚州的中央山谷地区。结果表明，OPTREF 模型有效地捕捉了研究区域三种典型土地覆盖类型（农田、草地和稀树草原）的区域植物特征和T/ET的季节变化，特别是农田替代种植系统引起的T/ET差异。重要的是，预测的T/ET值与之前的建模和经验报告非常一致；例如，耕地的T/ET范围为 [0.37-0.75]，草地为 [0.30-0.82]，稀树草原为 [0.39-0.77]。此外，还研究了 OPTREF 方法在使用蒸发干旱指数 ( EDI )跟踪农业干旱方面的适用性), 以及T/ET对干旱的生理反应分析。基于 OPTREF 的EDI始终如一地反映了植被健康指数记录和来自 NOAA 的干旱面积百分比数据所指示的相同干旱模式。此外，我们的研究结果表明，在极端干旱条件下，T/ET比率可能会增加，并且在耐旱植物覆盖的地区可能大约等于 1.0。总的来说，这种 OPTREF 方法为基于卫星的ET分区遥感和识别农业干旱提供了新颖、简单和通用性，特别是在气候快速变化的背景下。